Fix sampling in NuthKaab and set better default values for other classes

tests/test_coreg/test_affine.py

@@ -142,9 +142,8 @@ def test_coreg_example(self, verbose: bool = False) -> None:
         nuth_kaab.fit(self.ref, self.tba, inlier_mask=self.inlier_mask, verbose=verbose, random_state=42)
 
         # Check the output metadata is always the same
-        assert nuth_kaab._meta["offset_east_px"] == pytest.approx(-0.45061858808956284)
-        assert nuth_kaab._meta["offset_north_px"] == pytest.approx(-0.14225262689582596)
-        assert nuth_kaab._meta["vshift"] == pytest.approx(-1.987523471566405)
+        shifts = (nuth_kaab._meta["offset_east_px"], nuth_kaab._meta["offset_north_px"], nuth_kaab._meta["vshift"])
+        assert shifts == pytest.approx((-0.4648318628843316, -0.13376227526850593, -1.9860305501224076))
 
     def test_gradientdescending(self, subsample: int = 10000, inlier_mask: bool = True, verbose: bool = False) -> None:
         """

tests/test_examples.py

@@ -34,11 +34,11 @@ class TestExamples:
                 ddem,
                 np.array(
                     [
-                        -0.22833252,
-                        -0.82458496,
-                        0.18843079,
-                        1.0004578,
-                        -5.755249,
+                        0.0028076172,
+                        -0.7043457,
+                        0.14201355,
+                        1.1181335,
+                        -5.935547,
                     ],
                     dtype=np.float32,
                 ),

tests/test_spatialstats.py

@@ -494,7 +494,7 @@ def test_sample_multirange_variogram_default(self) -> None:
 
         # Check the variogram output is consistent for a random state
         df = xdem.spatialstats.sample_empirical_variogram(values=self.diff, subsample=10, random_state=42)
-        assert df["exp"][15] == pytest.approx(5.046060943603516, abs=1e-3)
+        assert df["exp"][15] == pytest.approx(5.11900520324707, abs=1e-3)
         assert df["lags"][15] == pytest.approx(5120)
         assert df["count"][15] == 5
         # With a single run, no error can be estimated
@@ -1286,7 +1286,7 @@ def test_patches_method_loop_quadrant(self) -> None:
         assert all(df.columns == ["nmad", "nb_indep_patches", "exact_areas", "areas"])
 
         # Check the sampling is fixed for a random state
-        assert df["nmad"][0] == pytest.approx(1.8530521253631773, abs=1e-3)
+        assert df["nmad"][0] == pytest.approx(1.8401465163449207, abs=1e-3)
         assert df["nb_indep_patches"][0] == 100
         assert df["exact_areas"][0] == pytest.approx(df["areas"][0], rel=0.2)
 
@@ -1295,7 +1295,6 @@ def test_patches_method_loop_quadrant(self) -> None:
 
         # Check the sampling is always fixed for a random state
         assert df_full["tile"].values[0] == "8_16"
-        assert df_full["nanmean"].values[0] == pytest.approx(0.3085488928369729, abs=1e-3)
 
         # Check that all counts respect the default minimum percentage of 80% valid pixels
         assert all(df_full["count"].values > 0.8 * np.max(df_full["count"].values))

xdem/coreg/affine.py

@@ -724,12 +724,14 @@ def _fit_func(
         if verbose:
             print("   Calculate slope and aspect")
 
-        valid_mask = np.logical_and.reduce((inlier_mask, np.isfinite(ref_dem), np.isfinite(tba_dem)))
+        slope_tan, aspect = _calculate_slope_and_aspect_nuthkaab(ref_dem)
+
+        valid_mask = np.logical_and.reduce(
+            (inlier_mask, np.isfinite(ref_dem), np.isfinite(tba_dem), np.isfinite(slope_tan))
+        )
         subsample_mask = self._get_subsample_on_valid_mask(valid_mask=valid_mask)
-        # TODO: Make this consistent with other subsampling once NK is updated (work on vector, not 2D with NaN)
-        ref_dem[~subsample_mask] = np.nan
 
-        slope_tan, aspect = _calculate_slope_and_aspect_nuthkaab(ref_dem)
+        ref_dem[~subsample_mask] = np.nan
 
         # Make index grids for the east and north dimensions
         east_grid = np.arange(ref_dem.shape[1])

xdem/coreg/biascorr.py

@@ -604,7 +604,7 @@ def __init__(
         fit_or_bin: Literal["bin_and_fit"] | Literal["fit"] | Literal["bin"] = "bin_and_fit",
         fit_func: Callable[..., NDArrayf] | Literal["norder_polynomial"] | Literal["nfreq_sumsin"] = "nfreq_sumsin",
         fit_optimizer: Callable[..., tuple[NDArrayf, Any]] = scipy.optimize.curve_fit,
-        bin_sizes: int | dict[str, int | Iterable[float]] = 10,
+        bin_sizes: int | dict[str, int | Iterable[float]] = 100,
         bin_statistic: Callable[[NDArrayf], np.floating[Any]] = np.nanmedian,
         bin_apply_method: Literal["linear"] | Literal["per_bin"] = "linear",
         subsample: float | int = 1.0,
@@ -810,7 +810,7 @@ def __init__(
         bin_sizes: int | dict[str, int | Iterable[float]] = 10,
         bin_statistic: Callable[[NDArrayf], np.floating[Any]] = np.nanmedian,
         bin_apply_method: Literal["linear"] | Literal["per_bin"] = "linear",
-        subsample: float | int = 1.0,
+        subsample: float | int = 5e5,
     ):
         """
         Instantiate a directional bias correction.